
Advertisement
Dirt 3 realistic driving mod
By
Levan Davitadze
, in Math and Physics
This topic is 1850 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.
If you intended to correct an error in the post then please contact us.
Recommended Posts
Levan Davitadze 111
Inertia tensor is my first thought, but that should be a 3x3 matrix and if I understand correctly it's only three values for each variation of the car. Maybe it's modeled as a similar shape and the inertia tensor can be represented as a vector for a box or something along those lines.
http://en.wikipedia.org/wiki/List_of_moments_of_inertia maybe "Solid cuboid of height h, width w, and depth d, and mass m" this is what i need?
Hodgman 51860
In PhysX, it (or the setMassSpaceInertiaTensor function) assumes that your 3x3 Inertia tensor matrix looks like
x 0 0 0 y 0 0 0 z
So instead of accepting a full 3x3 matrix, it just accepts a float3, which is the diagonal x/y/z values above.
I don't think Dirt uses PhysX, but it's another possibility to investigate.
maybe "Solid cuboid of height h, width w, and depth d, and mass m" this is what i need?
Yes, that formula is valid if you're representing the car's mass distribution as a solid cuboid / rectangular prism, which is very common in physics engines.
The I_{w}/I_{h}/I_{d} values in that math are the x/y/z values above  the inertia value for each axis.
Edited by HodgmanLevan Davitadze 111
thank you guys for help. only thing is confusingwhen i wrote dimensions in centimeters and weight in kgresult is too high values like (h 23 920 651), (w 22 897 005) and (d 5 283 854), if i divide it 10 000, tham maybe i get values that is "real". i can only guess what formula codemasters are using.
now it's clear how it works: there must be meters instead of centimeters :) now i have correct (i hope so) values (H 2392) (W 2289) (D 528). it's somehow works and car feels more realistic now. now i want to write down tyre wheel and engine inertias. i think there must be another formula for cylindrical object.
Edited by levmans
Advertisement
Inertia tensor is my first thought, but that should be a 3x3 matrix and if I understand correctly it's only three values for each variation of the car. Maybe it's modeled as a similar shape and the inertia tensor can be represented as a vector for a box or something along those lines.
Share this post
Link to post
Share on other sites